一种遗传算法与粒子群优化的多子群分层混合算法  

A multi-subgroup hierarchical hybrid of genetic algorithm and particle swarm optimization

作　　者：金敏[1] 鲁华祥[1] 

机构地区：[1]中国科学院半导体研究所人工神经网络实验室,北京100083

出　　处：《控制理论与应用》2013年第10期1231-1238,共8页Control Theory & Applications

基　　金：中国科学院战略性先导科技专项基金资助项目(XDA06020700);国家自然科学基金资助项目(61076014);江苏省高校自然科学研究基金资助项目(10KJA50042)

摘　　要：针对遗传算法全局搜索能力强和粒子群优化收敛速度快的特点,本文从种群个体组织结构上着手,进行优势互补,提出了一种遗传算法和粒子群优化的多子群分层混合算法(multi-subgroup hierarchical hybrid of genetic algorithm and particle swarm optimization,HGA–PSO).算法采用分层结构,底层由一系列的遗传算法子群组成,贡献算法的全局搜索能力;上层是由每个子群的最优个体组成的精英群,采用钳制了初始速度的粒子群算法进行精确局部搜索.文中分析论证了HGA–PSO算法具有全局收敛性,并采用7个典型高维Benchmark函数进行测试,实验结果显示该算法的优化性能显著优于其他测试算法.To make use of the strong global search ability of the genetic algorithm and the high convergence rate of the particle swarm optimization, we combine these two algorithms and propose a multi-subgroup hierarchical hybrid of genetic algorithm and particle swarm optimization （HGA-PSO）. This hybrid algorithm adopts a hierarchical structure; the base level is composed of a series of subgroups of Genetic algorithms, which provides the global search ability of the entire algorithm. The top level comprises all elite subgroups consisting of the best individual of each subgroup, which performs the accurate local search by using the particle swarm algorithm with cramped initial velocity. The global convergence analysis of HGA-PSO is given in this paper, and the performances of HGA-PSO have been evaluated through seven Benchmark functions. The experimental results show that the proposed method is superior to other related methods.

关 键 词：遗传算法 粒子群优化 分层混合算法 多子群 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]